| Message ID | 1433913284-18444-1-git-send-email-janusz.dziedzic@tieto.com (mailing list archive) |
|---|---|
| State | Accepted |
| Headers | show |
Hello, Thanks much for your help. However, I have looked at my driver code and did not find any implementations such as wiphy_ext_feature_set() or NL80211_EXT_FEATURE_VHT_IBSS. It seems not simple to apply your patch into my version of ath10k device driver. I am currently using backports-3.16-rc1-1 for ath10k device driver code in order to match the kernel version v3.16 for Ubuntu 14.04.02. Could you tell me the version of kernel and ath10k device driver of your test configuration? Or is there any way to apply your version of patch code into linux kernel v3.16? (I expect the version will be much more recent one than mine) Thanks, --Sangwoo Moon 2015-06-10 14:14? Janusz Dziedzic?(?) ??? ??: > Enable VHT support for IBSS, while mac80211/cfg80211 and > wpa_supplicant already support this. > > In my test env, ath10k 2x2 I get: > (udp) ath10k-1 >>>> ath10k-2 (server) - speed: 419 Mbits/sec > (tcp) ath10k-1 >>>> ath10k-2 (server) - speed: 404 Mbits/sec > > During tests I used wpa_supplicant (current version), which > already support IBSS VHT, and choose highest available BW. > > Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com> > --- > drivers/net/wireless/ath/ath10k/mac.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/drivers/net/wireless/ath/ath10k/mac.c > b/drivers/net/wireless/ath/ath10k/mac.c > index f4de717..14a8f25 100644 > --- a/drivers/net/wireless/ath/ath10k/mac.c > +++ b/drivers/net/wireless/ath/ath10k/mac.c > @@ -6949,6 +6949,8 @@ int ath10k_mac_register(struct ath10k *ar) > goto err_free; > } > > + wiphy_ext_feature_set(ar->hw->wiphy, NL80211_EXT_FEATURE_VHT_IBSS); > + > /* > * on LL hardware queues are managed entirely by the FW > * so we only advertise to mac we can do the queues thing
On 10 June 2015 at 20:04, Sangwoo Moon <swmoon@lanada.kaist.ac.kr> wrote: > Hello, > > Thanks much for your help. > However, I have looked at my driver code and > did not find any implementations such as wiphy_ext_feature_set() > or NL80211_EXT_FEATURE_VHT_IBSS. > It seems not simple to apply your patch into my version of ath10k device > driver. > I am currently using backports-3.16-rc1-1 for ath10k device driver code > in order to match the kernel version v3.16 for Ubuntu 14.04.02. > You can use newest backports with older kernel (new mac80211/cfg80211/drivers with old kernel). Or you can build and install new kernel for Ubuntu. I am using https://github.com/kvalo/ath BR Janusz > Could you tell me the version of kernel and ath10k device driver of your > test configuration? > Or is there any way to apply your version of patch code into linux kernel > v3.16? > (I expect the version will be much more recent one than mine) > > Thanks, > --Sangwoo Moon > > 2015-06-10 14:14? Janusz Dziedzic?(?) ??? ??: > >> Enable VHT support for IBSS, while mac80211/cfg80211 and >> wpa_supplicant already support this. >> >> In my test env, ath10k 2x2 I get: >> (udp) ath10k-1 >>>> ath10k-2 (server) - speed: 419 Mbits/sec >> (tcp) ath10k-1 >>>> ath10k-2 (server) - speed: 404 Mbits/sec >> >> During tests I used wpa_supplicant (current version), which >> already support IBSS VHT, and choose highest available BW. >> >> Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com> >> --- >> drivers/net/wireless/ath/ath10k/mac.c | 2 ++ >> 1 file changed, 2 insertions(+) >> >> diff --git a/drivers/net/wireless/ath/ath10k/mac.c >> b/drivers/net/wireless/ath/ath10k/mac.c >> index f4de717..14a8f25 100644 >> --- a/drivers/net/wireless/ath/ath10k/mac.c >> +++ b/drivers/net/wireless/ath/ath10k/mac.c >> @@ -6949,6 +6949,8 @@ int ath10k_mac_register(struct ath10k *ar) >> goto err_free; >> } >> >> + wiphy_ext_feature_set(ar->hw->wiphy, >> NL80211_EXT_FEATURE_VHT_IBSS); >> + >> /* >> * on LL hardware queues are managed entirely by the FW >> * so we only advertise to mac we can do the queues thing > > > -- > Sangwoo Moon, Ph.D. Candidate > Dept. of EE, KAIST, Korea > (Office) +82-42-350-5486, N1-#819 > (E-mail) swmoon at lanada.kaist.ac.kr > (URL) http://lanada.kaist.ac.kr > Gwahangno 335, Yuseong-Gu Daejeon, S. Korea 305-701 >
Hello, I have successfully installed the latest version of ath10k driver from backports. However, I have some difficulties on configuring wpa_supplicant for IBSS VHT mode. Looked at wpa_supplicant configuration file, but I couldn't found any option related to explicit setting VHT (though I've found disable_vht instead) Sorry for annoying questions, but can you help this out? At least you could let me know your configuration of wpa_supplicant. Thanks much, --Sangwoo Moon 2015-06-11 13:28? Janusz Dziedzic?(?) ??? ??: > On 10 June 2015 at 20:04, Sangwoo Moon <swmoon@lanada.kaist.ac.kr> > wrote: >> Hello, >> >> Thanks much for your help. >> However, I have looked at my driver code and >> did not find any implementations such as wiphy_ext_feature_set() >> or NL80211_EXT_FEATURE_VHT_IBSS. >> It seems not simple to apply your patch into my version of ath10k >> device >> driver. >> I am currently using backports-3.16-rc1-1 for ath10k device driver >> code >> in order to match the kernel version v3.16 for Ubuntu 14.04.02. >> > You can use newest backports with older kernel (new > mac80211/cfg80211/drivers with old kernel). > > Or you can build and install new kernel for Ubuntu. > I am using https://github.com/kvalo/ath > > BR > Janusz > >> Could you tell me the version of kernel and ath10k device driver of >> your >> test configuration? >> Or is there any way to apply your version of patch code into linux >> kernel >> v3.16? >> (I expect the version will be much more recent one than mine) >> >> Thanks, >> --Sangwoo Moon >> >> 2015-06-10 14:14? Janusz Dziedzic?(?) ??? ??: >> >>> Enable VHT support for IBSS, while mac80211/cfg80211 and >>> wpa_supplicant already support this. >>> >>> In my test env, ath10k 2x2 I get: >>> (udp) ath10k-1 >>>> ath10k-2 (server) - speed: 419 Mbits/sec >>> (tcp) ath10k-1 >>>> ath10k-2 (server) - speed: 404 Mbits/sec >>> >>> During tests I used wpa_supplicant (current version), which >>> already support IBSS VHT, and choose highest available BW. >>> >>> Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com> >>> --- >>> drivers/net/wireless/ath/ath10k/mac.c | 2 ++ >>> 1 file changed, 2 insertions(+) >>> >>> diff --git a/drivers/net/wireless/ath/ath10k/mac.c >>> b/drivers/net/wireless/ath/ath10k/mac.c >>> index f4de717..14a8f25 100644 >>> --- a/drivers/net/wireless/ath/ath10k/mac.c >>> +++ b/drivers/net/wireless/ath/ath10k/mac.c >>> @@ -6949,6 +6949,8 @@ int ath10k_mac_register(struct ath10k *ar) >>> goto err_free; >>> } >>> >>> + wiphy_ext_feature_set(ar->hw->wiphy, >>> NL80211_EXT_FEATURE_VHT_IBSS); >>> + >>> /* >>> * on LL hardware queues are managed entirely by the FW >>> * so we only advertise to mac we can do the queues thing >> >> >> -- >> Sangwoo Moon, Ph.D. Candidate >> Dept. of EE, KAIST, Korea >> (Office) +82-42-350-5486, N1-#819 >> (E-mail) swmoon at lanada.kaist.ac.kr >> (URL) http://lanada.kaist.ac.kr >> Gwahangno 335, Yuseong-Gu Daejeon, S. Korea 305-701 >>
2015-06-11 8:32 GMT+02:00 Sangwoo Moon <swmoon@lanada.kaist.ac.kr>: > Hello, > > I have successfully installed the latest version of ath10k driver from > backports. > > However, I have some difficulties on configuring wpa_supplicant for IBSS VHT > mode. > Looked at wpa_supplicant configuration file, but I couldn't found any option > related to explicit setting VHT (though I've found disable_vht instead) > > Sorry for annoying questions, but can you help this out? > At least you could let me know your configuration of wpa_supplicant. > supplicant by default setup wildest BW, no special parameter required for IBSS VHT network = { ssid="ibss" mode=1 key_mgmt=NONE frequency=5180 } Just use latest version of wpa_supplicant. next check you have 80MHz BW iw wlan0 info BR Janusz
On 06/10/2015 11:33 PM, Janusz Dziedzic wrote: > 2015-06-11 8:32 GMT+02:00 Sangwoo Moon <swmoon@lanada.kaist.ac.kr>: >> Hello, >> >> I have successfully installed the latest version of ath10k driver from >> backports. >> >> However, I have some difficulties on configuring wpa_supplicant for IBSS VHT >> mode. >> Looked at wpa_supplicant configuration file, but I couldn't found any option >> related to explicit setting VHT (though I've found disable_vht instead) >> >> Sorry for annoying questions, but can you help this out? >> At least you could let me know your configuration of wpa_supplicant. >> > supplicant by default setup wildest BW, no special parameter required > for IBSS VHT > > network = { > ssid="ibss" > mode=1 > key_mgmt=NONE > frequency=5180 > } > > Just use latest version of wpa_supplicant. > next check you have 80MHz BW > iw wlan0 info Make sure you compile supplicant with this option enabled: CONFIG_IBSS_RSN=y I think only particular firmware versions are doing to do IBSS as well. Thanks, Ben
Janusz Dziedzic <janusz.dziedzic@tieto.com> writes: > Enable VHT support for IBSS, while mac80211/cfg80211 and > wpa_supplicant already support this. > > In my test env, ath10k 2x2 I get: > (udp) ath10k-1 >>>> ath10k-2 (server) - speed: 419 Mbits/sec > (tcp) ath10k-1 >>>> ath10k-2 (server) - speed: 404 Mbits/sec What hw and firmware version? Do you expect this to work with both qca988x and qca6174? Oh, we should still add ATH10K_FW_FEATURE_NO_ADHOC so that we can disable ad-hoc on 10.x firmwares. Anyone have some spare time to cook up a patch? > During tests I used wpa_supplicant (current version) Better to say "current git version" (if that's what you mean).
On 11 June 2015 at 16:15, Kalle Valo <kvalo@qca.qualcomm.com> wrote: > Janusz Dziedzic <janusz.dziedzic@tieto.com> writes: > >> Enable VHT support for IBSS, while mac80211/cfg80211 and >> wpa_supplicant already support this. >> >> In my test env, ath10k 2x2 I get: >> (udp) ath10k-1 >>>> ath10k-2 (server) - speed: 419 Mbits/sec >> (tcp) ath10k-1 >>>> ath10k-2 (server) - speed: 404 Mbits/sec > > What hw and firmware version? Do you expect this to work with both > qca988x and qca6174? > Results are for qca9888x. But IBSS (VHT IBSS) works correctly also with qca6174 FW. > Oh, we should still add ATH10K_FW_FEATURE_NO_ADHOC so that we can > disable ad-hoc on 10.x firmwares. Anyone have some spare time to cook up > a patch? > I am not sure we need this, while we add ar->hw->wiphy->interface_modes |= BIT(NL80211_IFTYPE_ADHOC) only for main and tlv version. Seems we already cover 10.x case correctly. >> During tests I used wpa_supplicant (current version) > > Better to say "current git version" (if that's what you mean). > > -- > Kalle Valo
On 06/11/2015 10:12 PM, Janusz Dziedzic wrote: > On 11 June 2015 at 16:15, Kalle Valo <kvalo@qca.qualcomm.com> wrote: >> Janusz Dziedzic <janusz.dziedzic@tieto.com> writes: >> >>> Enable VHT support for IBSS, while mac80211/cfg80211 and >>> wpa_supplicant already support this. >>> >>> In my test env, ath10k 2x2 I get: >>> (udp) ath10k-1 >>>> ath10k-2 (server) - speed: 419 Mbits/sec >>> (tcp) ath10k-1 >>>> ath10k-2 (server) - speed: 404 Mbits/sec >> >> What hw and firmware version? Do you expect this to work with both >> qca988x and qca6174? >> > Results are for qca9888x. > But IBSS (VHT IBSS) works correctly also with qca6174 FW. > >> Oh, we should still add ATH10K_FW_FEATURE_NO_ADHOC so that we can >> disable ad-hoc on 10.x firmwares. Anyone have some spare time to cook up >> a patch? >> > I am not sure we need this, while we add > ar->hw->wiphy->interface_modes |= BIT(NL80211_IFTYPE_ADHOC) > only for main and tlv version. Seems we already cover 10.x case correctly. If you give me some way to advertise my CT 10.1 firmware supports ADHOC mode in the upstream driver then there will be one more firmware variant that works with IBSS in stock kernels.... Maybe a feature flag 'ATH10K_FW_FEATURE_SUPPORTS_ADHOC' ? You can special case older firmware that is known to support ADHOC but has no feature flag for it. And new firmware could add the feature flag as needed. Thanks, Ben
Janusz Dziedzic <janusz.dziedzic@tieto.com> writes: > On 11 June 2015 at 16:15, Kalle Valo <kvalo@qca.qualcomm.com> wrote: >> Janusz Dziedzic <janusz.dziedzic@tieto.com> writes: >> >>> Enable VHT support for IBSS, while mac80211/cfg80211 and >>> wpa_supplicant already support this. >>> >>> In my test env, ath10k 2x2 I get: >>> (udp) ath10k-1 >>>> ath10k-2 (server) - speed: 419 Mbits/sec >>> (tcp) ath10k-1 >>>> ath10k-2 (server) - speed: 404 Mbits/sec >> >> What hw and firmware version? Do you expect this to work with both >> qca988x and qca6174? > > Results are for qca9888x. > But IBSS (VHT IBSS) works correctly also with qca6174 FW. Ok, thanks. I added this info to the commit log in the pending branch. >> Oh, we should still add ATH10K_FW_FEATURE_NO_ADHOC so that we can >> disable ad-hoc on 10.x firmwares. Anyone have some spare time to cook up >> a patch? >> > I am not sure we need this, while we add > ar->hw->wiphy->interface_modes |= BIT(NL80211_IFTYPE_ADHOC) > only for main and tlv version. Seems we already cover 10.x case correctly. Ah, I had forgotten that. Good that we have that.
Janusz Dziedzic <janusz.dziedzic@tieto.com> writes: > Enable VHT support for IBSS, while mac80211/cfg80211 and > wpa_supplicant already support this. > > In my test env, ath10k 2x2 I get: > (udp) ath10k-1 >>>> ath10k-2 (server) - speed: 419 Mbits/sec > (tcp) ath10k-1 >>>> ath10k-2 (server) - speed: 404 Mbits/sec > > During tests I used wpa_supplicant (current version), which > already support IBSS VHT, and choose highest available BW. > > Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com> Thanks, applied.
diff --git a/drivers/net/wireless/ath/ath10k/mac.c b/drivers/net/wireless/ath/ath10k/mac.c index f4de717..14a8f25 100644 --- a/drivers/net/wireless/ath/ath10k/mac.c +++ b/drivers/net/wireless/ath/ath10k/mac.c @@ -6949,6 +6949,8 @@ int ath10k_mac_register(struct ath10k *ar) goto err_free; } + wiphy_ext_feature_set(ar->hw->wiphy, NL80211_EXT_FEATURE_VHT_IBSS); + /* * on LL hardware queues are managed entirely by the FW * so we only advertise to mac we can do the queues thing
Enable VHT support for IBSS, while mac80211/cfg80211 and wpa_supplicant already support this. In my test env, ath10k 2x2 I get: (udp) ath10k-1 >>>> ath10k-2 (server) - speed: 419 Mbits/sec (tcp) ath10k-1 >>>> ath10k-2 (server) - speed: 404 Mbits/sec During tests I used wpa_supplicant (current version), which already support IBSS VHT, and choose highest available BW. Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com> --- drivers/net/wireless/ath/ath10k/mac.c | 2 ++ 1 file changed, 2 insertions(+)